Large-Scale Real-Time Optimization In The Delayed Coking Production Process

The lightening of heavy oil has always been a very important part in the petroleum refining and chemical industry.The delayed coking process has become an important means of handling heavy oil due to its low production requirements.The heavy oil is first heated in a heating furnace to reach a preset temperature,and then enters the coke tower for cracking and condensation reaction,and the generated coke is left in the tower to wait for decoking operation.The remaining oil and gas enter the fractionation tower and absorption stabilization device to generate subsequent products.However,due to the operating characteristics of its semi-intermittent production,it is extremely difficult to control the delayed coking production process.Although some refineries use advanced control(APC),they are all divided and conquered.Even if each part reaches the local optimum,because of the existence of the coupling of the controllers,it puts.a greater pressure on the operator and the device itself,resulting in an increase in eneigy consumption and affecting yield.For the delayed coking production process,the reduction of energy consumption and the improvement of yield are its main research priorities.There are generally two ideas for reference.On the one hand,the device itself can be changed by understanding the process,but this method is time-consuming and expensive,and it is difficult to implement.On the other hand,it can only change the operating point of the production to complete the corresponding requirements.Correlation integration method is a steady-state optimization method for solving multi-input multi-output continuous real-time optimization problems.It can achieve operation tuning of industrial processes without mathematical modeling and has strong anti-interference.The research content of this article is mainly the following:First of all,for the problem of large energy consumption of delayed coking,relevant research is done on the tube heating furnace.Combined with the material and heat balance conditions of the tube heating furnace,a state space equation is constructed,and a particle filtering-based joint method is used to estimate the predicted value of the model in real time,and the real-time estimated value is tuned by using the relevant integration method to reduce energy consumption.Secondly,the low yield of delayed coking products leads to the problem of low overall economic benefits.Based on the original correlation integration low overall economic benefits.Based on the original correlation integration method,the augmented Lagrangian method is used to constrain the tuning variables to achieve the optimization of the industrial process operation under the condition that the constraint input range is met.The feasibility of the method is proved,and simulation experiments are carried out on related tools to verify the rationality of the method.Finally,based on the principle of adding constraint-related integrals,the multi-variable optimization control system is improved,and the interface of tuning variable constraints is added.The included modules are briefly introduced and the feasibility and effectiveness are illustrated by relevant experiments.